Authors’ reply to “Labor analgesia and postpartum depression: Methodological and statistical issues”

Dear Editor, We thank the authors for raising several pertinent issues regarding our recently published paper.[1] We agree with some of their observations but not all. First, regarding the “ethical” concerns flagged by the authors; the Institute Ethics Committee approved the study protocol. The protocol was also registered at the Clinical Trials Registry-India, where the ethical aspects were again declared. Specifically, as clearly mentioned in the methods, the control group participants, who underwent normal vaginal delivery received the standard treatment as per the hospital protocol. Hence there was no ethical violation. Just to give more details, pain management in treatment as usual group (TAU) was intravenous analgesia as per our institutional protocol/practice that included paracetamol 1 gm i.v. to all parturient, along with tramadol 100 mg i.v. as needed. They were explained about the analgesia and risks and benefits associated and a written informed consent was obtained. Second, regarding the administration of the baseline instruments and collection of baseline data, these were administered and collected at the early phase of first stage of labor, in between the bouts of labor pain, and when the participants could cooperate. To the best of our knowledge, there is no contraindication for such procedures in this situation. Further, because the participants were randomized, the risk of bias was minimal. Third, we agree that Edinburgh Postnatal Depression Scale (EPDS) is a validated screening tool for postpartum depression (PPD), which is a psychiatric diagnosis that can only be confirmed by a clinician. Hence, we argue, it is better to use EPDS score as a marker of PPD symptoms severity, (which was our primary objective) rather than use it for a clinical diagnosis of PPD syndrome. We do not agree that this was a “failed RCT.” Indeed, the mean EPDS score at 6 weeks postpartum (the primary outcome) was 3.48 ± 1.61 for the index group and 6.97 ± 1.52 for the control group [Table 2 of the cited paper]. The mean EPDS score at 6 weeks postpartum when compared, among the two groups was statistically significant (P < 0.001), with a substantive effect size of 1.49. Fourth, EPDS is measure of postnatal depression; it is invalid for “baseline” assessment. The 3-day EPDS score is more for capturing “postpartum blues” (a much commoner but transient state, as reflected in the higher 3-day EPDS scores in both the groups), not to serve as baseline score. The baseline depressive assessment was conducted using PHQ-9 during the prenatal assessment, which showed that there was no significant difference between the two groups (Group 1: 0.84 ± 0.86; Group 2: 1.10 ± 0.79; P = 0.143; Table 2 of the paper). Finally, regarding the three finer statistical points raised by the authors, we agree that more refined analysis can further improve the results but do not agree with the aspersion that our results are invalid. Choosing the independent variables in a multiple regression model from those with significant results in bivariate analysis is a well-established procedure even if less than perfect. The purpose of this exercise (as in Table 3 of the cited paper) was to estimate the strength of prediction of these selected variables by observing the standardized beta coefficient, along with their statistical significance. The State-Trait Anxiety Inventory scores do show variance [Table 2 of the cited paper]. In summary, we again thank the authors for raising valuable issues. However, we feel that, while there is always room for improvement and correction, our study did not violate principles of research ethics, was conducted with a reasonable degree of methodological and statistical rigor, and the findings are interpretable, valid, and clinically meaningful. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.